Study of financial access to health services for the poor in Cambodia

Transcription

1 PHASE 2 REPORT Study of financial access to health services for the poor in Cambodia Phase 2: In-depth analysis of selected case studies Dr Peter Leslie Annear (RMIT University) in association with Maryam Bigdeli (WHO Cambodia) Ros Chun Eang (MOH Cambodia) Prof. Paul James (RMIT University) For the Ministry of Health, WHO, AusAID and RMIT University 30 June 2007 Phnom Penh

2 This is the Final Report for Phase 2 of the Study of Access to Health Services for the Poor in Cambodia supported by AusAID, WHO Cambodia, RMIT University (Melbourne) and the Ministry of Health Cambodia. The Final Report fulfils the requirements of Milestone 6 under AusAID Contract 13732, Variation Order No. 2 of 31 July 2006, Clause 6.1 (i). Funding for the research was agreed under the terms of the Contract (CON 13732) signed in 2005 between AusAID and RMIT University for implementation of the study project under the title The cost, benefit and sustainability of equity funding for health: evidence from Cambodia. Terms for Phase 2 of the study were formalized by the Contract Variation to the Scope of Services finalized in July 2006.

3 Contents Glossary Executive summary iii v 1 Background Health financing schemes Research methods Research questions Study sites Data collection methods Selection of respondents Triangulation Ethics Approval Data collection The data collection team Training workshops Data collection activities Data entry and analysis Profile of interviewees and informants Findings: analysis and discussion Coverage of pro-poor schemes Access to public health facilities Utilization at referral hospitals Contracting of health services User fees and other revenues Additional patient charges Exemptions for the poor Staff behaviour Fairness and stigma Participation and empowerment Debts and asset sales Contrasts and comparisons Reporting and dissemination of findings Conclusions...43 Figures Figure 1. User Fees, Contracting, Health Equity Funding, and Community-Based Health Insurance...2 Figure 2. Operational Health Districts and Study Sites...5 Figure 3. Study site profiles...5 Figure 4. Per cent of IPD discharges by gender: PPMH January-December Figure 5. Exit interviewees: patients with HEF or SKY...18 Figure 6. Exit interviewees: patients who did not attend the facility prior to having HEF or SKY...19 Figure 7. IPD discharges, Phnom Penh MH...21 Figure 8. OPD consultations, Phnom Penh MH...22 Figure 9. IPD discharges, Ang Roka RH...23 Figure 10. OPD consultations Ang Roka RH...23 Figure 11. Bed occupancy rate, Phnom Penh MH...24 i

4 Figure 12. Bed occupancy rate, Ang Roka RH...24 Figure 13. Number of HEF-supported IPD admissions, Phnom Penh MH...25 Figure 14. Fee-paying and HEF-assisted patients, Ang Roka RH Figure 15. Number of SKY members enrolled, Ang Roka OD...26 Figure 16. Main sources of hospital revenue: Phnom Penh MH...29 Figure 17. Main sources of hospital revenue: Ang Roka RH...29 Figure 18. Patient knowledge of user fees, HEF and SKY...30 Figure 19. Patients who paid more in addition to user fees...31 Figure 20. Proportion of all exit interviewees receiving fee-exemptions...33 Figure 21. Exit interviewees: debts and asset sales for health costs...37 Figure 22. Comparison of HEF and SKY between Phnom Penh and Ang Roka...41 Annexes A. Ethics approval 49 B. Data collection instruments 51 C. Data collection schedules 61 D. Training workshop summaries 63 E. Respondents and informants 69 F. Site profiles 72 G. Exit interview analysis 78 ii

5 Glossary ADB AFH AMDA AusAID BHE BFH BOR BTC CAAFW CAS CBHI CDHS CEDAC CFDS CHHRA CIDA CMS CON DFID DPHI EF EI FGD EU FMFA GTZ GRET HEF HC HCP HFC HIS HNI HSSP HU INGO ID IPD JFPR KII LNGO MH MHD MOH MOP MSF NGO Asian Development Bank Action for Health Asian Medical Doctors Association Australian Agency for International Development Bureau of Health Economics of the DPHI Buddhists for Health Bed occupancy rate Belgian Technical Cooperation Cambodian Association for Assistance to Families and Widows Centre for Advanced Studies, Phnom Penh Community-Based Health Insurance Cambodia Demographic and Health Survey Centre d'etude et de Développement Agricole Cambodgien Cambodian Family Development Services Cambodian Health and Human Rights Alliance Canadian International Development Agency Central Medical Supply (MOH) Contracting Department for International Development (UK) Department of Planning and Health Information (MOH) Equity Fund Exit interviews Focus group discussions European Union French Ministry of Foreign Affairs Gesellschaft fuer Technische Zusammenarbeit (German Technical Cooperation) Groupe de Recherche et d Echanges Technologiques (SKY manager) Health Equity Funding Health Centre Health Coverage Plan Health Financing Charter Health Information System (MOH) Health Net International Health Sector Support Project (funded by the ADB, WB, DFID, JFPR) Health Unlimited International NGO Identification In-patient department Japan Fund for Poverty Reduction Key informant interviews Local NGO Municipal Hospital (Phnom Penh) Municipal Health Department (Phnom Penh) Ministry of Health Ministry of Planning Médecins Sans Frontières Non-government organisation iii

6 ODO OPD OD PHD PH PPMH RH RMIT SHI SCA SKY SRC UHP URC UNFPA USAID USG VHSG WB WHO OD administrative office Out-patient department Operational District Provincial Health Department (MOH) Provincial Hospital Phnom Penh Municipal Hospital Referral Hospital Royal Melbourne Institute of Technology (Australia) Social Health Insurance Save the Children Australia Health for Our Families (phonetic for Khmer translation) Swiss Red Cross Urban Health Project University Research Company United Nations Fund for Population United States Agency for International Development Urban Sector Group Village Health Support Group World Bank World Health Organization iv

7 Executive summary Introduction: The Study of Financial Access to Health Services for the Poor in Cambodia has been carried out in two phases. Phase 1 of the Study (completed in April 2006), made a comprehensive national analysis access to health services for the poor across all health districts in Cambodia where Contracting, Health Equity Funding (HEF) and/or Community-Based Health Insurance (CBHI) schemes had been introduced. Phase 2 builds on the Phase 1 findings through the analysis of two in-depth case studies in one urban and one rural location Phnom Penh (the capital city) and Ang Roka (a rural district in Takeo Province two hours drive south-east of the capital). The case study approach was designed to validate the findings of the Phase 1 research and to provide richer and more detailed quantitative and qualitative data on the operation and impact of the pro-poor schemes, focussing on the demand-side issues of access to health services. Phase 2 collected data on facility utilization and revenues, the impact of user fees on service users, and the implementation of Contracting, HEF and CBHI schemes. Of the two sites, Phnom Penh had official user fees, HEF and CBHI (through the SKY insurance scheme) while Ang Roka had these three schemes as well as Contracting. The main interest of the Access Study is the impact of these various schemes separately and together on access to public health services for the poor. In practice, user fees are universal and apply to all patients unless they qualify for formal (means-tested and funded) exemptions or informal exemptions (offered by health staff to poor people when required). Contracting covers the management and administration of all public health services in the selected district (such as Ang Roka). HEF is targeted on the poor and is implemented in Phnom Penh and Ang Roka, following an objective means test administered through household surveys; it is available automatically to all those identified through the means test. And SKY is a voluntary community-based health insurance scheme relying on self-selection through the purchase of premiums, and is targeted on the not-so-poor (workers in the formal and informal sectors). Methodology: Phase 2 of the Access Study addressed a number of key questions related to these issues: Are user-fee systems applied appropriately? Do user fees exclude the poor from health services? Do exemptions, HEF and CBHI schemes provide increased access to services? Is the quality of care the same as for those who pay fees? What are the public perceptions of user fees, HEF, CBHI and health services? Are there significant rural-urban differences? The research used a number of different quantitative and qualitative methods of data collection and triangulated the data to establish valid findings. Data collection was carried out following an analysis of the published literature on relevant issues and a documentary analysis of all relevant reports on user fees, Contracting, HEF and CBHI in Cambodia. The literature review and documentary analysis built on and updated the work carried out in Phase 1 of the Study. v

8 Routine quantitative data on referral hospital utilization and revenues were collected for Phase 2 through the Ministry of Health from the official Health Information System. Routine quantitative data on HEF attendances at facilities and SKY enrolment were collected from the NGO scheme providers. Primary data was collected through different quantitative and qualitative methods, which included: Structured patient exit interviews at the referral hospitals and all health centres at the two study sites. Focus group discussions with HEF beneficiaries, SKY beneficiaries, and nonbeneficiaries of any scheme at the two study sites. Key-informant interviews with MOH provincial and district health managers and staff and local and international NGO providers of Contracting, HEF and SKY programs at the two study sites. The study sites were chosen against criteria that included: An urban site. A rural site. A site that includes all schemes (User Fees, Contracting, HEF and CBHI). The site had reliable data on utilization, exemptions and revenues for at least the last five years. The site had reliable data on the implementation of HEF and/or CBHI arrangements for at least the last two years. For the primary data collection, the aim was to survey a sufficient number of respondents through key informant interviews, focus group discussions and patient exit interviews to get a complete and accurate view of the situation. Sites for the data collection included the Municipal Health Department (MHD), the Phnom Penh Municipal Hospital and all seven health centres in the Municipality supported by HEF schemes. In Ang Roka, the data collection sites included the Provincial Health Department (PHD), Provincial Hospital, OD district health office, the Referral Hospital and all nine health centres in the district. In total, primary data was collected from the following number of respondents: Patient exit interviews: A total of 679 patients from all facilities Focus group discussions: A total of 12 focus group discussions involving a total of 166 participants Key informant interviews: A total of 42 key informants equally distributed across the two sites. Ethics approval was granted by the Research Ethics Committee of the Ministry of Health, a plain language Respondent Information Sheet was prepared and issued to data collectors, who were trained in its use at Training Workshop sessions, and a Statement of Informed Consent was prepared to guide data collectors and to inform respondents about the purposes and methods of the study. Data collection, data entry and preliminary analysis were carried out with the assistance of staff from the Department of Planning and Health Information (MOH) and the Centre vi

9 for Advanced Studies (CAS) under the supervision of the Chief Investigator. Data collection was completed in Phnom Penh from 9 to 13 October and in Ang Roka from 20 to 23 November. Workshop training was provided for all data collectors prior to data collection. The data collection team comprised the following staff: Findings: The Chief Investigator (Peter Annear) The WHO health financing advisor (Maryam Bigdeli) The director and four staff of the MOH Health Economics Bureau (Ros Chun Eang, Ngin Seilaphiang, Thor Bony, Kim Lunsithan, Phum Phat) One research associate (Men Rithy Chean) Eleven researchers from the Centre for Advanced Studies A research assistant (Phy Sopheada). In general, the evidence suggests that HEF and SKY provided access for many people who did not previously attend public health facilities, some because they could not afford it and for some because they used alternative providers. For more than half of patients attending the facilities at the time of data collection, the hospital or health centre was their first point of treatment. Among the most important findings of the research were: HEF is the only mechanism that provides access to public health services for those people who previously could not attend because they did not have the money to meet the user fees and associated costs. SKY-CBHI provided a mechanism to encourage the greater use of public health services by patients who had previously used alternate services, including more expensive state hospitals and private providers. Supply-side financing and management mechanisms such as user fees and Contracting work best to improve the utilization of public health services when used in combination with demand-side measures such as HEF and CBHI. The effectiveness of HEF and CBHI may depend on the context and the conditions of implementation and management. Together, HEF and SKY provided coverage for about 30% of patients attending the facilities in Phnom Penh and Ang Roka at the time of data collection. The coverage provided by HEF and SKY in each location was influenced by the length of time each had been implemented and to the strength of management in the implementation of the schemes, including the completeness of pre-identification processes for HEF. HEF had been implemented since 1999 in Phnom Penh but only since April 2005 in Ang Roka. SKY insurance had been available in Phnom Penh only since December 2005 and in Ang Roka from June In Phnom Penh, HEF was available at the Municipal Hospital and seven health centres; in Ang Roka it was available (formally) only at the Referral Hospital (and informally through health centres). In Phnom Penh SKY insurance was provided through one health centre and the Municipal Hospital and in Ang Roka at all nine health centres and the Referral Hospital. vii

10 There was a feeling among respondents that the poor relied increasingly on public services for their health care and could not afford private providers. Utilization of the referral hospitals in both locations had increased steadily in recent years, aided by the contribution of the HEF and SKY schemes. It appeared that HEF had helped to overcome financial barriers facing the poor, and that SKY had encouraged those with a little more money to attend public facilities rather than alternative providers. The steady growth of IPD and OPD attendances evident at the two referral hospitals has not in general been seen at comparable MOH facilities that do not enjoy external support or have the HEF or SKY schemes. In the first nine months after its introduction in Ang Roka, the HEF provider reported that more than 60% of RH patients were support by HEF. There was also evidence that health centres had performed more strongly in recent years, relieving the burden of OPD attendances at the referral hospitals and strengthening the referral system. Respondents believed that the provision of Contracting services in Ang Roka OD had had a positive impact on staff incentives and health facility management, ensuring in particular that services were able to operate 24 hours a day. While Contracting was valued, there was also a strong view among local health administrators that it was time to hand over the responsibility for implementing the Contracting arrangements to MOH personnel (in a similar manner to the arrangements at Takeo Provincial Hospital). User fees together with HEF and SKY have become an important source of revenue for facilities on top of government funding for infrastructure costs, staff and drug supplies. Patients considered the fee schedules to be reasonable and fair to most people, though still a barrier to the very poor. At facilities, normal exemptions were offered to the poor but were regarded as a drain on facility revenues. Consequently, the combination of user fees for those who could afford them and funded exemptions, like HEF, was seen as the best alternative to guarantee access for the poor. The evidence from patients and health administrators suggests that, together, Contracting, HEF and SKY had helped to limit and control the practice of under-the-table charges at public health facilities, and had helped to improve staff behaviour towards patients. The contractual arrangements and performance agreements that accompanied these schemes, as well as the monitoring and supervision, had worked to provide harsh penalties when such practices were discovered. According to patients, however, under-the-table charges had not been eliminated. In general, community knowledge and understanding of the user fee system, of HEF and of SKY is limited. While health centres (particularly those with external NGO support) conduct outreach activities and community structures (like village chiefs or Village Health Support Groups) are used to publicise the nature of these schemes, in many cases almost half of the patients interviewed had not been well aware of them. The provision of HEF and SKY are regarded by patients and be health administrators as fair, with few complaints about false inclusions and exclusions. However, particularly in Ang Roka, it appeared that people who qualified for HEF may not have yet received it. In general, HEF and SKY patients were satisfied with the treatment they had received from the health staff, though there were still complaints about the behaviour of some doctors and nurses. There was no evident stigma associated with HEF beneficiaries, and HEF membership was valued. viii

11 The degree to which communities had been involved in the setting of user fees (as is required by health financing regulations), and the degree to which patients had been empowered in their dealings with the health system, were both limited. Health administrators testified that they had in some cases carried out consultations with community leaders or local administrators in setting fees, but real community participation was lacking and patients felt they had not been fully consulted. In Phnom Penh, the HEF implementer (USG) had been more active in working within the communities and provided community liaison workers, but less had been done in this respect in Ang Roka. Borrowing and asset sales remained common, even among HEF and SKY beneficiaries, but it appears that the schemes may increase the opportunity for discretionary use of money set aside for health care. Few rural-urban differences were evident, though pre-identification for HEF seemed more effective in Phnom Penh, and in Ang Roka there was much more widespread uptake of SKY, reflecting the longer period of implementation there. Demographically and financially, the urban setting was more divers and more complex than the rural. The level of debts and asset sales in Ang Roka were greater than in Phnom Penh, perhaps reflecting the higher levels of poverty. Conclusions: Almost universally in Phnom Penh and in Ang Roka the main benefit derived from the availability of the HEF and SKY schemes was regarded as the reduced costs of treatment and improved access. This was evident from all sources of data. Among the significant conclusions to be drawn from the Phase 2 study are: HEF and CBHI can provide coverage for a significant proportion of the population and protect the poor HEF provides access to health services for the poor HEF and SKY help to increase the use of public facilities HEF and SKY help to reduce the problem of under-the-table charges but do not yet fully empower patients HEF and CBHI increase the scope for discretionary use of OOP payments for health care HEF is regarded as fair with no feelings of stigma felt by HEF beneficiaries. User fees, Contracting, HEF and CBHI work best in combination. In general, the findings from exit interviews, focus group discussions and key informant interviews were consistent, and indicated that HEF and SKY worked to improve access to health services for the poor and the near-poor. It appears that these health financing schemes, along with Contracting procedures, have acted to improve the behaviour of the health staff towards patients and to make services more responsive to the needs of the poor. HEF, SKY and Contracting all appear to help achieve more accountability on the side of the service providers. ix

12 1 Background The Study of Financial Access to Health Services for the Poor assessed the national health and poverty relationship in Cambodia, with particular reference to financial access to health services by the poor in two locations, one urban and one rural, and evaluated the impact of existing health financing and social protection schemes that act to alleviate the burden of health care costs on the poor. The research considered issues related to four major health financing and pro-poor schemes User Fees, Contracting, Health Equity Funding (HEF) and Community-Based Health Insurance (CBHI) (see Figure 1 below.). Phase 1 of the Study of Financial Access to Health Services for the Poor, completed in April 2006, made a comprehensive national analysis of access to health services for the poor across all health districts in Cambodia where Contracting, HEF and/or CBHI schemes had been introduced. The full report of Phase 1 findings is available at and at Phase 2 of the Access Study, completed in June and 2007, built on the research completed in Phase 1 through in-depth case studies in two selected health districts to identify the impact of the various financing schemes on access to health services for the poor. The case study approach is designed to validate and deepen the findings of the Phase 1 research by in-depth analysis of facility utilization and patient attitudes in the selected health districts. Phase 1 looked for common trends in utilization, fee-exemptions and revenues among all health districts with the identified financing schemes. Phase 2 investigated the conditions of access to health services in Phnom Penh Municipality and in the Ang Roka Health Operational District, where these schemes have been in place for some time. In brief, Phase 1 of the research indicated that: Contracting increases access to health services by strengthening management and quality of service but does not specifically target the poor; HEF specifically targets the poor and increases their access to health services; CBHI targets those living just above the poverty line who may otherwise become impoverished by health costs. Phase 2 of the research focused principally on demand-side issues in access to health services. Of the four health-financing and pro-poor schemes included in the study, User Fees and Contracting are clearly supply-side initiatives designed to improve the management and financing of public health services. HEF and CBHI act on the demand side to remove financial barriers to access to public health services. All four schemes act in different ways to improve the quality of service delivery. Phase 2 commenced on 1 July 2006 and was completed on 30 June Data collection was completed by 31 December Partners in the Phase 2 study included AusAID, WHO Cambodia, Ministry of Health (MOH) Cambodia and the Globalism Institute at RMIT University (Melbourne). The Study Team comprised Dr Peter Annear (RMIT University), Dr Lo Veasna Kiry, Mr Ros Chun Eang, Ms Thor Bony (and staff from the Department of Planning and Health Information, MOH), and Ms Maryam Bigdeli (WHO Health Financing Officer), with support from Mr Men Rithy Chean and researchers from the Centre for Advanced Studies (CAS). Translation and research assistance was provided by Mr Phy Sopheada and Mr Chap Prem. 1

13 Figure 1. User Fees, Contracting, Health Equity Funding, and Community-Based Health Insurance User Fees Contracting Health Equity Funding Community-Based Health Insurance Official user fees at public health facilities were introduced in Cambodia through the 1996 Health Financing Charter. The Charter certifies the imposition of official fees according to an agreed schedule at affordable rates following consultation with the community. The initiative to implement fees remains with hospitals and health centers. Health facilities must apply to the MOH for permission to implement fees, based on minimum standards of service delivery. Denotes a scheme in which all government health services at health district level (primary-level care at health centres and secondary-level care at district referral hospitals) are managed and delivered by a non-government operator working under contract to the MOH, using MOH staff with performance agreements, and funded in Cambodia through the ADB/WB Health Sector Support Project (in 11 health districts). A third-party payer scheme for indigent patients in which a fund is managed at district level by a local agent (usually NGO), supervised by an international NGO, and funded by donors (or in some cases through community collections). The poor are identified at or prior to the point of service and receive free care at the health facility. The facility then receives reimbursement monthly directly from the fund for services provided to the poor. May eventually become tax-funded. Local-level insurance schemes funded by user premiums and managed commonly by an international or local NGO. The insurer contracts public health facilities to provide approved health services. Contracted facilities receive a monthly payment through capitation or case payment for services provided to CBHI beneficiaries. May eventually be included under the umbrella of the anticipated tax-funded social health insurance scheme. 2 Health financing schemes The Cambodian public health system is organised into 76 different operational health districts (ODs), each serving a population of roughly 100,000 to 200,000, each with a district referral hospital, and each with primary-care health centres. Under its broader program of health sector reform, the Ministry of Health began a pilot program for the implementation of a system of official User Fee schemes in a number of public health facilities. The 1996 National Health Financing Charter extended the opportunity to approved public health facilities to levy and collect User Fees according to an agreed scale formed in consultation with local communities. The aim was to generate extra revenues at facility level and to create a managed environment for improving service quality. Consequently, 99% of the revenues generated were to be retained at facility level and channelled back into staff incentives (50%) and operational costs (49%). Subsequently these proportions were amended to 60% and 39%. The intention was to reduce the actual costs of health care to the patient, enhance staff motivation, suppress unofficial fees, improve transparency, improve quality of care and improve access to public health services for the majority of the population, including the poor. The User Fees system included the right of facilities to grant fee exemptions to very poor patients, though in practice the exemptions system has worked poorly and has covered only a proportion of the genuinely poor. The term Contracting denotes a scheme in which government health services within one health district are managed and delivered by a non-government operator working under contract to the Ministry of Health (MOH), employing MOH staff under workperformance agreements. First trialled after 1998, Contracting now operates in 11 of Cambodia s 76 health districts under the Health Sector Support Project (HSSP), with funding from the Asian Development Bank, World Bank, UK Department for International Development and the UNFPA (a new phase of the HSSP project will begin in 2008). Contracting is designed to provide stronger management of service delivery and improved quality of care at government facilities. Previous evidence from the contracting 2

14 pilots suggests that this is a cost-effective approach. One of the aims of Contracting is to improve health service delivery for the poor, and the scheme is targeted on poorer, more remote districts. Introduced in numerous locations around Cambodia since 1999, HEF has been independently sponsored by a number of non-government organisations (NGOs) working to support the government health system. In principle, the HEF is a targeted socialprotection scheme in which a fund is established as a third-party payer for indigent patients attending government health facilities. There is not a single, uniform model of HEF among the numerous schemes now in operation, but they do all share basic characteristics. The HEF generally operates at the district referral hospital, with financing provided mainly by donors, though funds may also be collected from the community. Generally, management is provided by an international NGO, which sub-contracts fund administration to a local NGO. The eligible poor are screened and means-tested at or prior to the point of service (known respectively as post-screening and preidentification ). The facility maintains a record of user-fee exemptions given to identified HEF patients and is reimbursed monthly from the fund. The local fund managers also act as advocates for the identified poor at the health facility and monitor the quality of care provided. The main aims of HEF are to reduce financial barriers, facilitate access to priority public health services, promote public health service utilization, reduce out-ofpocket health care costs, protect the poor from catastrophic expenditures, improve the quality of the public health services, and increase the accountability of service providers by giving a voice to service users. Also introduced first in 1999, CBHI has grown more slowly than HEF and is currently provided by two independent NGOs in a small number of health districts. While there is little previous literature on this form of micro-insurance, the potential for growth is well recognized in Cambodia. CBHI is a voluntary, locally-based program provided by independent insurance funds for services provided at nominated government health facilities (in-patient and out-patient). Premiums are less than $2 per month per family maximum. Field agents are employed at village level to market policies, administer monthly premium collection at family level, and negotiate with facilities. CBHI operates as a third-party purchaser of health services, generally through a capitation payment made monthly in advance to the facility. CBHI therefore works to increase utilization of public health services, provides an additional source of revenue to facilities, and may act as a voice for patients, with financial leverage. While CBHI is a non-profit scheme it must achieve financial sustainability. It therefore requires a satisfactory quality of care from the facility and a clear referral system at primary-care level. In Cambodia, the MOH provides the regulatory framework for HEF and CBHI, which may one day be included within a broader social health insurance system. 3 Research methods Phase 2 of the research uses case-study methodology to verify and deepen the analysis conducted in Phase 1. Two sites were chosen as case studies, an urban site in the capital Phnom Penh and a rural site in the health operational district of Ang Roka, two hours drive south-east of the capital. The case-study approach makes use of a number of different quantitative and qualitative data collection techniques to gather information from a broad range of sources on access to health services for the poor. These are discussed below after first describing the main research questions and identifying the 3

15 study sites. The different methods of data collection are used to triangulate the data in order to strengthen the reliability and validity of the findings. 3.1 Research questions The purpose of the Phase 2 research was to gather detailed, in-depth information on demand-side issues related to access to health services and to test the conclusions drawn from data collected in Phase 1. The main research questions in Phase 2 were: Are user-fee systems applied appropriately? Do user fees exclude the poor from health services? Do exemptions, HEF and CBHI schemes provide increased access to services? Is the quality of care the same as for those who pay fees? What are the public perceptions of user fees, HEF, CBHI and health services? Are there significant rural-urban differences? 3.2 Study sites The study sites in Phnom Penh and Ang Roka were chosen against a number of criteria designed to guarantee both the availability of reliable data and the coverage of the main supply-side and demand-side financing schemes that influence access to services for the poor. The criteria were: One urban and one rural site. A site that includes all schemes (User Fees, Contracting, HEF and CBHI). The site had reliable data on utilization, exemptions and revenues for at least the last five years. The site had reliable data on the implementation of HEF and/or CBHI arrangements for at least the last two years. Phnom Penh: Phnom Penh is a diverse urban location where both HEF and CBHI schemes are implemented through the Municipal Hospital. In Phnom Penh, the study focused on the Municipal Hospital, which acts as the Referral Hospital for four ODs within the city, and seven health centers where HEF is provided. These health centers serve a population including a number of identified squatter settlements where poor and dispossessed families commonly reside. CBHI is offered under the SKY ( Health for Our Families ) scheme with health services provided through a dedicated health centre (SKY Health Centre) located at the Phnom Penh Municipal Hospital (PPMH), including referral to the PPMH. No Contracting of health services was implemented at any of the ODs or facilities at the Phnom Penh site. Ang Roka: Ang Roka is a rural location and one of only two locations in country where Contracting, HEF and CBHI are implemented together through the Ang Roka Referral Hospital and health centres. 4

16 In Ang Roka, the study focused on the Referral Hospital (RH) and all nine health centres in the Operational District. Data were also collected from key informants at the Takeo Provincial Hospital, to which HEF and CBHI patients may be referred. HEF exemptions are funded by the provider (Health Net International/Action For Health) only at the Ang Roka RH and not at health centres; however, Swiss Red Cross (SRC) and other organizations reimburse health centres for fee exemptions they offer to HEF card holders at health centers. CBHI is offered also under the SKY scheme for health services provided at the RH and all nine HCs in the OD. Contracting of health services, covering the RH and all HCs, has been provided through two different organizations since Figure 2 illustrates all 76 health Operational Districts in Cambodia, and identifies the two study sites. Annex F includes a detailed profile or each of the study sites, which are summarized briefly in Figure 3 below. Figure 2. Operational Health Districts and Study Sites Phnom Penh Ang Roka Figure 3. Study site profiles Phnom Penh Ang Roka Population >1.3m. (in 4 ODs) ~130,000 (in 1 OD) Poverty ~12% (>22 squatter settlements) ~30% (rural) User fees began - Referral Hospital Health centers HEF began 1999 (now at PPMH and 7 HCs, April 2005 (RH only) incl. 6 settlements) SKY began December 2005 ( now >300 policy holders) June 2001(now >3000 policy holders) Contracting began None January 1999 (with AMDA; through SRC from 2004) 5

17 3.3 Data collection methods In general, the Phase 2 study collected data from both referral hospitals and health centres and from health administrators and scheme implementers. However, in general, the main impact of user fees and relief provided for the poor through HEF and CBHI is at referral hospital level. Health costs are greater and constraints on access for the poor are more evident at RH level than at HC level, where health costs are low. In Ang Roka, the formal HEF scheme was implemented only at the referral hospital. Routine HIS data was collected only at the referral hospital level. A wide range of data were collected for the Study using three main approaches: 1. A documentary analysis of all relevant reports and published literature on Contracting, HEF and CBHI in Cambodia; 2. A quantitative analysis of routine data of health facility utilization and revenues; and 3. Analysis of primary data collected from patient exit interviews, focus group discussions and key informants. Documentary analysis A complete documentary analysis of all relevant reports and published literature on Contracting, HEF and CBHI in Cambodia was completed for Phase 1 of the Study and a database of all relevant literature was constructed. For Phase 2, the literature analysis and the Study database were both updated, particularly for new materials published in 2006 and The documentary analysis provided the basis for focussing on the most important domains in primary data collection. Quantitative analysis For the study sites Phnom Penh and Ang Roka the database of quantitative indicators constructed in Phase 1 was revised and updated. For Phase 2, routine quantitative data on referral hospital utilization and revenues were collected from the official Health Information System of the Ministry of Health. The main concern for quantitative data analysis was to measure levels of facility utilization and revenues and the impact of the different health-financing and pro-poor schemes. The quantitative data was collected at referral hospital level in the two study sites using a number of key indicators, including: Monthly number of in-patient attending the hospital Monthly number of out-patients seeking consultations at the hospital Utilization of hospital capacity measured by the bed occupancy rate Hospital revenues from government budget, user fees and funded exemptions (HEF and SKY) The level of informal or other exemptions (non-hef or SKY) granted by the hospitals. Additional routine quantitative data was collected from the international and local NGOs working as the managers and implementers of the HEF and SKY schemes covering: The level of HEF population coverage within the municipality or district The level of utilization of the hospital by HEF beneficiaries The number of enrolled SKY beneficiaries. 6

18 Primary data collection and analysis Primary data was collected for the Phase 2 study through different quantitative and qualitative methods, which included: Structured patient exit interviews at the referral hospitals and all health centres at the two study sites. Focus group discussions with HEF beneficiaries, SKY beneficiaries, and nonbeneficiaries of any scheme at the two study sites. Key-informant interviews with MOH provincial and district health managers and staff and local and international NGO providers of Contracting, HEF and SKY programs at the two study sites. Data collection instruments Survey instruments were prepared for each of the various methods of data collection, as follows (copies of all data collection instruments are attached in Annex B): The questionnaire for collecting quantitative data on health facility utilization and revenues used for Phase 1 was used again for the Phase 2 case studies, updating and cross checking the data collected in the first phase. New data was collected from the MOH HIS. The range of data collected was extended from the previous three years used in Phase 1 back to the year 2000 or the earliest year for which data were available. A coded questionnaire was used for exit interviews with provision for openended replies to a number of closed questions. The additional replies were post-coded and all data were translated and entered into SBSS for analysis. The exit interviews were treated as a valid non-probability sample of patients attending the Referral Hospital and Health Centres on the days of data collection. Frequencies were calculated to assess common patient responses and the results were statistically tabulated. Separate open-ended question guides were prepared for each of the main areas for focus group discussions: HEF beneficiaries, CBHI beneficiaries, Nonbeneficiaries. The focus group discussions were conducted in Khmer and were led by researchers from the Centre for Advanced Studies. Each FGD was digitally recorded, transcribed by staff from the MOH and CAS, and translated into English for analysis. A question guide was prepared for each of the main key-informant groups: MOH administrators and staff, Contracting managers, managers of HEF schemes, and managers of CBHI schemes. The instruments were used as a guide for semi-structured, open-ended interviews. The interviews were conducted by the research team in Khmer for all MOH staff and Khmer-speaking scheme managers and in English for managers of international NGOs. The interviews were digitally recorded, transcribed and (where necessary) translated into English for analysis. The analysis was conducted using identified key themes to sort responses. 7

19 3.4 Selection of respondents Data were collected from the main health providers, official records and from respondents selected as exit interviewees, focus group attendees and key informants. Quantitative data were collected from the MOH Health Information System for each of the referral hospitals (Phnom Penh and Ang Roka). For the primary data collection, the aim was to survey a sufficient number of respondents through key informant interviews, focus group discussions and exit interviews to get a complete and accurate view of the situation. Sites for primary data collection included the Municipal Health Department (MHD), the Phnom Penh Municipal Hospital and all seven health centres in the Municipality supported by HEF schemes. In Ang Roka, the sites for primary data collection included the Provincial Health Department (PHD), Provincial Hospital, OD office, Referral Hospital and all nine health centres in the OD. A detailed account of the number and composition of all exit interviews, focus groups discussions and key informant interviews is provided in Annex E. Data was collected as follows: Patient Exit Interviews The aim was to interview all daily visitors to the referral hospital and all health centres at each site on the days allocated to data collection (four days at each referral hospital and one day at each health centre in both locations). The number of respondents to be targeted was calculated on the basis of an estimation of the actual average daily number of IPD and OPD patients attending the facilities. Exit interviews conducted: Total number n = 679 (PPMH and seven HCs; ARRH and nine HCs) Phnom Penh n = 429 (PPMH = 200; total of seven HCs = 229) Ang Roka n = 250 (ARRH = 100; total of nine HCs = 150) Focus Group Discussions Respondents were chosen from three different categories at each site: HEF beneficiaries, SKY beneficiaries, and non-beneficiaries of either scheme. To reduce bias and increase coverage, two focus groups were formed for each of these three categories at each site. The selection of participants for focus groups was arranged by the NGO providers under criteria established by the research team. In each group the participants were to include a representative number of participants by sex, by age and by occupation designed as far as possible to reflect the communities from which they came. Focus Group Discussions: A total of 12 different focus groups (6 Phnom Penh, 6 Ang Roka) Including 2 focus groups in each location for HEF beneficiaries, CBHI beneficiaries, and non-beneficiaries Total number of participants: 166 (78 in Phnom Penh, 88 in Ang Roka, an average of 14 per group) 8

20 In Phnom Penh, FGD participants were chosen as follows: HEF beneficiaries: Two groups arranged by field workers from USG including registered HEF beneficiaries from a more remote site adjacent to Anlong Kngan Health Centre in Group 1 and from two sites more central to the city near to Boeungkak and 7 Makara health centres in Group 2. SKY beneficiaries: Two groups arranged by field workers from GRET, Group 1 made up of formal-sector workers in full-time paid employment and Group 2 made up of informal-sector workers with no regular formal employment. Non-beneficiaries: Two groups also arranged by field workers from USG including people without HEF or SKY coverage from the more remote site at Anlong Kngan in Group 1 and from the two sites more central to the city at Boeungkak and 7 Makara in Group 2. In Ang Roka, FGD participants were chosen as follows: HEF beneficiaries: One group arranged by field workers from the HEF provider (Action For Health AFH) at the Ang Roka RH to assess conditions in the OD and one group arranged by the HEF provider at the Takeo Provincial Hospital (Cambodia Family Development Services CFDS) to assess the strength of the referral system. SKY beneficiaries: Two groups arranged by field workers from GRET (the French NGO managing the scheme), one at the Tram Kak health centre and one at the Prey Chour health centre. Non-beneficiaries: Two groups also arranged by field workers from SRC including people without HEF or SKY coverage, one at the more remote village at Trapeang Chouk and one at the Takeo Provincial Hospital. Key Informant Interviews Key Informant Interviews were held with Provincial and District Health Department staff and with hospital and health centre directors in both locations and with managers from the NGO service providers. All key leaders of these organizations were interviewed. Key Informant Interviews: In Phnom Penh: 22 key informant interviews (including MHD directors, PPMH directors, OD directors, HC directors and NGO managers and providers) In Ang Roka: 20 key informant interviews (including PHD directors, Provincial Hospital directors, RH directors, OD directors, HC directors and NGO managers and providers) All interviews with Cambodian staff were conducted in the Cambodian language. These interviews were carried out by Central Ministry of Health staff under the supervision of the Chief Investigator. Key Informant Interviews with directors and managers of all foreign and local NGOs providing Contracting, HEF or CBHI services were conducted by the Chief Investigator in English or in Khmer through a translator. All interviews were electronically recorded and transcribed and where necessary were translated into English. Transcription and translation were carried out by researchers from the Ministry of Health and Research Assistants under the supervision of the Chief Investigator. 9

21 3.5 Triangulation This was not a sample survey producing statistical results but a case study based on the use of a number of different techniques of quantitative and qualitative data collection. To establish the validity and reliability of the findings, the data from all sources the quantitative data from health facility records as well as the primary data from key informant interviews, exit interviews with facility users, and focus group discussions were triangulated. Only those conclusions that were reliably substantiated by this process of cross-checking and validation are reported. 3.6 Ethics Approval Approval from the Research Ethics Committee of the Ministry of Health for the research involving human subjects (key-informant interviews, focus group discussions, and exit interviews) was sought and granted. All interview and focus-group respondents were provided with a plain-language explanation of the research and gave their informed consent to participation in the research. Those not willing to participate (particularly in exit interviews) were recorded as non-respondents. A copy of the ethics approval is attached in Annex A. A plain language Respondent Information Sheet was prepared and issued to data collectors, who were trained in its use at Training Workshop sessions. A Statement of Informed Consent was also prepared to guide data collectors in the requirements for ensuring that all respondents were well informed, clear about the purposes and methods of the study and the use of data, and confident that their anonymity would be guaranteed. Additionally, with the support of the Secretary of State for Health at the central MOH, notices were issued to all MOH units participating in the study at Provincial and OD level ensuring them that their cooperation was authorized by the Ministry and that they were free to participate openly in responding to the data collection activities. Within the context of the Cambodian civil service this arrangement is common and accepted practice and necessary to the implementation of research projects like this one. As some respondents were civil servants and many EI and FGD respondents were likely to have limited literacy skills, there was a requirement on the part of data collectors that they explain clearly to respondents the nature of the research and that they get their verbal agreement to participate. On the EI questionnaire, for example, the first question was: Do you agree to participate in this interview? A similar process was followed in the selection of FGD participants. 4 Data collection The research team spent five days in Phnom Penh and four days in Ang Roka OD to collect the data, with additional time spent collecting key-informant interviews from NGO personnel based in Phnom Penh and the HIS quantitative data. Copies of the schedules for data collection activities in Phnom Penh and Ang Roka are attached in Annex C. 10